SPSS Statistics

SPSS Statistics

Your hub for statistical analysis, data management, and data documentation. Connect, learn, and share with your peers! 


#Analytics
#SPSSStatistics
#Analyticstools
 View Only

  • 1.  Bonferroni in Pairwise Comparisons

    Posted 05/06/24 10:11 AM

    Hello, I tried to compare my data to see the mean difference. As shown in the table, the mean differences are the same (1.113) for both comparisons, but another table does not show significant difference. Can you explain to me why this happened? or is there any criteria for this? thank you.
    table 1

    Pairwise Comparisons

    Measure:   comp_check 

    (I) task

    (J) task

    Mean Difference (I-J)

    Std. Error

    Sig.a

    95% Confidence Interval for Differencea

    Lower Bound

    Upper Bound

    1

    2

    .667

    .494

    .199

    -.394

    1.727

    2

    1

    -.667

    .494

    .199

    -1.727

    .394

    Based on estimated marginal means

    a. Adjustment for multiple comparisons: Bonferroni.


    table 2

    Pairwise Comparisons

    Measure:   conf_check 

    (I) task

    (J) task

    Mean Difference (I-J)

    Std. Error

    Sig.a

    95% Confidence Interval for Differencea

    Lower Bound

    Upper Bound

    1

    2

    1.133

    1.687

    .513

    -2.485

    4.752

    2

    1

    -1.133

    1.687

    .513

    -4.752

    2.485

    Based on estimated marginal means

    a. Adjustment for multiple comparisons: Bonferroni.


    ------------------------------
    kiromil baroroh
    ------------------------------


  • 2.  RE: Bonferroni in Pairwise Comparisons

    Posted 05/06/24 10:26 AM
    These are different variables.  Why would you expect the tests to give the same results?

    Second, you are only testing one post-hoc hypothesis, so there is no Bonferroni correction.  Bonferroni adjusts the sig level as alpha/number of hypotheses, so it is just alpha with one test.

    You may be thinking that the number of hypotheses adjusts across variables, but it is actually adjusted for the number of groups.  If you do want a multiple-testing correction across variables, you can use the STATS PADJUST (Analyse > Descriptive Statistics > Calculate Adjusted p Values) extension command.  If you don't have that installed, you can get it via Extensions > Extension Hub.

    --





  • 3.  RE: Bonferroni in Pairwise Comparisons

    Posted 05/06/24 11:07 AM

    Hello. First of all, thank you for the answers.

    Secondly, I am sorry, I happened to send the wrong table. Here I put the complete tabulated result. 

    Comparison

    Comprehension check

    Confirmation Check

    Clarification Request

    Task 1 

    vs.

    Task 2

    .667

    1.133

    7.200*

    Task 1 

    vs.

    Task 3

    1.133*

    1.533

    1.400

    Task 2

    vs.

    Task 3

    .467

    2.667

    5.800*

    *. The mean difference is significant at the 0.05 level.

    Previously, I analyzed these variables by using Repeated Measures ANOVA in SPSS to see the mean differences. As shown in the table, there are the same result numbers of mean difference (1.133) for different comparisons (task 1 vs 2) and (task 1 vs 3). Although they showed the same mean difference, only one shows significance. Also, 2.667 is higher than 1.533 etc. but why does it not show a significant difference? Are there any criteria for this? Or maybe it's because I didn't analyze it correctly.
    I am looking forward to receiving your explanation. Thank you.



    ------------------------------
    kiromil baroroh
    ------------------------------

    Original Message


  • 4.  RE: Bonferroni in Pairwise Comparisons

    Posted 05/06/24 01:00 PM
    The significant test depends not only on the mean difference but on the standard deviations.  Since there is a different group involved, that is likely to be different for the two tests even if the mean difference is the same (at least to the displayed precision).

    --



    Original Message


  • 5.  RE: Bonferroni in Pairwise Comparisons

    Posted 05/06/24 11:30 PM
      |   view attached

    I read somewhere on the internet which stated that if the value of standard error is lower than the mean difference, it would show a significant difference, and vice versa. Is it correct?
    However, some of my data does not show as stated above (you can see my attached file)



    ------------------------------
    kiromil baroroh
    ------------------------------

    Attachment(s)

    docx
    analisis SPSS - Kiromil (2).docx   41 KB 1 version
    Original Message


  • 6.  RE: Bonferroni in Pairwise Comparisons

    Posted 05/07/24 03:08 PM
    You can find the formulas in the SPSS Algorithms manual available via Help > Doc in PDF Format.

    --



    Original Message


  • 7.  RE: Bonferroni in Pairwise Comparisons

    Posted 05/08/24 02:09 AM

    Alright, thank you so much, Sir.



    ------------------------------
    kiromil baroroh
    ------------------------------

    Original Message


Related Content